The Exchanges

Q Wait, how does that work? So, so that's a very, it's a good specific example. So I guess what you would do is Issue the purchase order for all of the material before you commit to a price to the customer? Like how do you avoid the terror?

A Yeah, exactly. It's timing it all out. So, um, when, uh, uh, just to compare this again to how it works today is you'd only, when you, when you say, Hey, we're the EPC, we're going to give you a performance guarantee or some sort of fixed firm price when you go to buy this or fund this project with like a notice to proceed or final investment decision. At that point, you just, you literally don't even know All your bulks, like, so your, all your commodity equipment, your, the steel that you need, the, the wire that you need, etc. Um, we will know all of that at this point in time because we're not 30% definition, we're a hundred percent engineering definition. And because we have this entire equipment list at this point in time, um, we know either how to say, hey, let's purchase everything at the, at the same day that the project gets funded, let's go either purchase everything, or let's design hedges from a cost perspective so that we're not exposed to some sort of volatility or risk.

AI assessment note: “let's purchase everything at the, at the same day that the project gets funded”

Q Can you articulate a little bit, like, what is distinct about this from the, you know, that EPCs have been using engineering software that is super mature for, for a long time. Maybe that's the problem, but, like, what's distinct about what you just described from what you could just get off the shelf to design a project?

A Most of the software in this industry really hasn't Hasn't changed much in the last like maybe 20 years. Um, and I, again, kind of always come back to the incentive structure. There's no real incentive structure for an EPC to want better software. It's the same reason your law firm doesn't necessarily want like AI software to help them accelerate, you know, and reduce their billables. So the way these softwares work now is that there's a bunch of kind of distributed software solutions or, um, or maybe decentralized solutions. So you'll have a CAD tool that you use to actually like do the designs. You'll do almost all of your kind of rough engineering or hand calcs in Excel. Um, you'll use dedicated simulation software for some sort of fluid design or structural analysis or whatever that might be. And then, um, and then you use another, another tool to like review those, those PDFs, you know, uh, uh, with a tool to put red lines on them and go through design cycles and then push to drafters and, um, and then other tools to manage, you know, how do you reach out to your vendors? And anyways, there, there, there's all these different solutions that, uh, software solutions that you're using in order to, um, like kind of quote unquote, like do engineering. And, um, Um, what, what we built, uh, kind of from day one was just a modern version of the software. So, uh, we put all those d…

AI assessment note: “we put all those different disparate tools into one underlying platform”

Q to hedge them every single time if they weren't going to do that. Like the, and not to mention labor availability, labor costs, and all this other stuff. So is it because they just can't wear that kind of a risk, the like time risk? Or is your experience that there's a different reason why it ended up being Cost plus or fixed firm with all sorts of change orders?

A No, it's, I mean, definitely to Steelman, it's, it's because it's ultra complex to build these projects. There is hundreds of people that'll be involved in how the thing gets designed, um, that manages the entire supply chain to get it built, and then manages all the, like, boots on the ground of the project that's actually doing the building. So there is tons of complexity, and the reason that these contracts are structured in this Um, inherently flexible way is because, um, it's really hard to know up front how much exactly is the thing going to cost. So, um, yeah, it's, I mean, it, it, it certainly is like evolved this way because when you, when you think about these projects, and I mean, the small version of these are like a hundred million dollars. The, the, you know, the more normal version of these is many hundreds of millions, and then the, you know, the bigger ones are in the, certainly in the, in the, uh, media billions of, of, of, of costs. So, um, because you have all that risk, the contracts effectively are this risk Big, like, you know, risk mitigation strategy. And, um, and that, that, the easy way for the EPC not to take any risk is just to say we're gonna get a margin on top of what it costs us to build the thing. But the, the problem is, um, that, that there's no, there's no motivation then to, like, make the thing cost less. If anything, there's a motivation …

AI assessment note: “reason that these contracts are structured in this Um, inherently flexible way is because”

Q let's talk about how to use technology, which is, I think, what you're alluding to here, right? Like, this is an area where my suspicion is generally it's not the first category of AI adoption, or at least it hasn't been historically. Where do you see the biggest opportunity? Like, where are the step function improvements that one can achieve by leveraging modern technology, be it AI or something else?

A Yeah, um, So, uh, we have built a lot of technology to help accelerate that pre-construction phase. So this is basically the E in lots of the procurement management side of things. And, um, we have a software and, uh, kind of like an AI native version of this that helps accelerate the chemical design, mechanical, electrical, civil, structural, um, controls. And in doing that, we basically are, um, building a platform that allows us to remove that risk, but at a very, very low marginal cost. So that helps you accelerate through the design process, not just saving time, but exploring a much wider search base and getting to much greater levels of definition before you actually have to freeze designs. Um, and that has a big impact to the previous point, right? It's just like removing uncertainty, um, you know, reducing costs from through, through optimization.

AI assessment note: “we have built a lot of technology to help accelerate that pre-construction phase”

Q What, you mentioned data centers, and I, I think you, you guys put that in as one of the categories you're most interested in, in, uh, in your announcement when you guys unveiled. Is that because we're building a lot of data centers? Or is that because there's like something specific to data centers, data center Construction and engineering that makes it especially attractive here.

A Or is there something else that you would describe as like the, the perfect prototypical use case for this in the early days is X. So, uh, these projects have, you know, a handful of different, a handful of different reasons of why you would apply this technology. Um, in the data center use case specifically, speed is really an important, um, lever for them. And so you can take, you know, instead of the design portion taking, call it six to nine months of your critical path. Uh, we can massively collapse that period of time. And that's really, that's really important for, for these data center projects to, um, kind of accelerate how quickly we can actually build them. Um, the same piece though, with the fact that, you know, if you right now budget six to nine, um, months for, for your design portion, almost none of that, you're gonna start to figure out, well, how do we optimize that design or do some sort of value engineering to save costs or, um, to design for long lead items or to design for constructability. So, really strong value prop in these data center projects. Um, as we, as we end up kind of increasing the capacity that we have on our team, though, it'll make sense to pick up, um, basically many of these types of large industrial projects, um, of which we've already pushed the kind of technology to be able to demonstrate, and, um, and it'll end up being, like, really…

AI assessment note: “in the data center use case specifically, speed is really an important, um, lever”

Q that vary. And then construction, which is sort of a function. It's an outcropping of, of how the engineering occurs, but certainly is also just driven by, um, By construction process and, and labor and so on. So as you think about, like, where is there an opportunity to deliver cost reduction and also cost certainty? Like, how much of it is the E versus the P versus the C?

A Yeah, totally. Okay, so those are two separate things, right? Like, so let's say, um, cost certainty versus, like, an overall reduction in cost. Um, on the certainty side, uh, everything it's Really kind of planned out or teased out in the engineering portion, right? And so generally again, life cycle of these projects are, um, you know, you start with some conceptual design, you understand maybe plus or -50%, how much the thing's gonna cost. Um, then for the most part, you dig into this like front end engineering design and you will get to totally changes per project, but say you get to like plus or -10% of cost. And that gives you enough confidence to go to a lender, an actual bank and say, hey, let's underwrite this project. We think you're gonna get the IRR you need. Um, and then we move forward. But in the, in the vast majority of projects, when you go to final investment decision and you have plus or -10% estimates, you're only maybe 30% of the way through your engineering. 70%, you still have to go do. And you just, it doesn't make sense to do it up front because it takes a lot of time and it costs a lot of money and it's a lot easier to finance that, um, with the actual lender's money rather than the development capital. So, um, the, the certainty that you have going into a project before you actually take the money from the bank To go build it is actually pretty low. L…

AI assessment note: “On the certainty side, everything it's really kind of planned out in the engineering portion”

Q Well, it's actually the, the problem with solar is that the, the module costs have fallen much faster than the installed costs have, right? Like we haven't solved it on the construction and, and all the other commodity materials and stuff.

A Yeah, and that's because all progress happens with this, like, you know, highly iterative process, and this, like, continuing improvement process that we get out of manufacturing a thing. But when we move into projects world, everything's a snowflake. It's like, you know, everything's end of one. Even a project that we're building, you know, we built lots of refineries. That next refinery that we build is still an end of one project. And because it's approached in that way, you don't really get any sort of benefits of learning through that process. And, you know, by definition, each project's going to be unique. It's a different piece of land. Um, but, uh, yeah, and again, maybe this is a, a, a tangent, but, um, that's like another big problem in the industry is everything's done from scratch. So even if you're gonna go design a new, um, you know, a slightly different data center or solar project or something else of that sort, uh, Every tweak or change that you make effectively reintroduces the entire redesign of the process. And so, yes, there's some stuff that gets lifted, but the overall amount of work that happens is not the kind of delta of the change. It is like a significant rehaul of work. Um, and it keeps, it just keeps us from like learning and iterating to improve, uh, how something gets designed, not just in, in terms of saving the engineering costs, which is actua…

AI assessment note: “that's because all progress happens with this... But when we move into projects world”

Q Can you go a little deeper on that? Like, what does that actually mean to reduce uncertainty using software? Like, what are you actually doing?

A Yeah. Um, it's just, you know, if you start a project today at like zero, you know, it's like completely uncertain, right? Um, and as you start specifying or defining the, um, bringing definition to that project, which is the process of engineering it, you start to see more, you start to remove risk, right? You say, hey, you know, we can actually size a pump that will solve that, you know, moving a fluid from, from point A to point B. Or, um, a, you know, actually we can design a tank like this to hold this amount of volume. And once you start providing that level of definition, you, you remove a lot of the kind of ambiguity, but there's still tons and tons that goes into, you know, the thousands of, of documents that you'll, you'll kind of generate to design a plant fully. Um, so this is what we kind of built. Like it is a, um, a platform that our engineers use. I think the important distinction here is this is not like a software technology that we sell. It's like we use this as an internal tool ourself. Um, So our engineers have this, uh, kind of AI platform that has all of their data in one place, all the kind of requirements about the project, all of the, um, potential vendors that could be used, and, um, they use it to basically accelerate that design process really significantly. We haven't, like, redefined how you're engineering these projects. Um, we've, uh, kind of au…

AI assessment note: “our engineers have this, uh, kind of AI platform that has all of their data”